General Discussion
Related: Editorials & Other Articles, Issue Forums, Alliance Forums, Region ForumsAn assumption and a warning about AI
My Assumption: Garbage in, garbage out.
The Warning: Beware. Be Careful.
If an expert can refute or clarify, please feel free.

no_hypocrisy
(53,135 posts)Dont believe everything you read.
customerserviceguy
(25,402 posts)People will believe anything that comes out of the ass of a computer.
JCMach1
(28,989 posts)Gemini and Notebook LM spent several hours last night tutoring my son in Chemistry for his test today. His teacher isn't the greatest so he first used the podcast to give him a 411 lecture from his study and work notes. He then created flashcards and practice quizzes. AI creates these in just a few minutes.
He also knows to walk back through when using AI for tutoring for the occasional hiccup. You should always doublecheck math and formulas anyway, so not a bad habit.
AI in this case is actually better than the teacher for him.
If I had the same tools in HS... Just saying.
So is someone going to tell me my kid should not use the tools that help him get past the limitations of his teacher (you have no choice of who you get)and his socioeconomic status (we cannot afford expensive tutors).
AI is largely just a tool. Don't buy into the hype that is used to both sell it and by others to ban it.
WarGamer
(17,921 posts)While your GIGO observation is true, here is how it works.
I'll use Google Gemini for example because I know it best.
Google has partnered with libraries, universities, museums and cultural/governmental institutions for their data.
So if you ask a question about George Washington, NO you're not going to find info posted by some rando on Reddit.
Sources for Gemini include:
Digital Public Library of America
Europeana
HathiTrust Digital Library
Internet Archive
Library of Congress
Project Guttenberg
Smithsonian Libraries and Archives
New York Public Library Digital Collections.
And Google works with smaller groups like the Archive of the Egyptian Museum Cairo (AEMC) for everything they have including 50,000 photos, manuscripts and plans.
Google uses filters and ranking systems to judge quality of sources so don't worry about gigo, particularly with established topics like mathematics, engineering, history and philosophy.
AI sometimes tends to have difficulty with current events, pop culture and such things
As a university professor, we are still trying to get a handle on how to use AI to promote student learning and also help ourselves with our academic research. The main AI tools have gotten better in regard to the reliability and validity of the output, so long as the prompts are well-constructed.
The caveat is that writing effective prompts and reviewing the output requires some degree of knowledge or expertise in the field.
My concern overall is that companies are in such a rush to use AI that they may not have thought through how best to use it or the ethical considerations of using it. As with any technology, it can be used for good or ill.
Trust_Reality
(2,241 posts)haele
(14,613 posts)Where it falls down is the ability to assess that data on more than its search keywords.
Say I ask it an archeological assessment question like "What did the Neolithic people use their hill forts for?" and put in several common defensive, environmental, or trading criteria along with the catalogue of what has found at Neolithic Hill forts.
Is AI going to notice that some structures that were called Neolithic Hill forts do not have evidence of wells being dug inside fort structures, so maybe, they're something else? Whereas a human being will go "wait a minute,what is missing?" when looking over the site.
I wouldn't use AI to write my answer for me. Just maybe act as my library...but that's only if I know it's not also scraping Social Media.
Tarc
(10,595 posts)